Snowball methodology: Novel application of an age-old research strategy to capture elusive study populations

Introduction: Hidden population groups pose a challenge to epidemiologists. Snowball sampling, also known as referral sampling, can be employed to track elusive or stigmatized population groups, making it ideal for the study of any group that does not readily identify itself, such as commercial sex workers or illicit drug users. In this case, snowball methodology was employed to study long term health and socioeconomic outcomes of child development programs in the Dominican Republic.

Methods: Staff at thirty child development program centers each identified a “first-level” subject: an adult who previously was a beneficiary of the program ten years or more ago. The research team contacted the first-level individuals, interviewed them, and asked them for referrals to anybody they knew from their program cohort group. The research team continued down the trail of referrals, to no more than the 4th level individual. Data were analyzed using Epi Info.

Results: On average, each first-level contact was able to give 2-4 referrals. Of an estimated 15,000 potential study participants, 500 names were obtained, and 123 individuals were able to be contacted and interviewed. Female and male respondents were equal in number. Health and socioeconomic data were collected and presented elsewhere.

Conclusions: Snowball methodology is a viable strategy for epidemiologic research on elusive study groups that are otherwise unable to be studied. Disadvantages include: absence of a clearly defined study population, and the presence of inherent selection biases. This study demonstrated the applicability of snowball methodology to outcomes research on child development programs.